correcting video delay introduced by NeatVideo ?

questions about practical use of Neat Video, examples of use
Post Reply
Fudoh
Posts: 7
Joined: Mon Jul 24, 2006 3:10 pm

correcting video delay introduced by NeatVideo ?

Post by Fudoh »

Hi,

I'm using Neatvideo on some VHS rips I've done some years back.

When I open a video in VirtualDub I can see that the right (output) windows lags approx. 1 frame behind the left (input) window.

Also - when using VDub's frameserver I get a warning that Neatvideo introduces a video lag.

Since I don't want to re-decode, re-edit and re-encode the audio of the video (AC3 stream), I need to re-adjust the video after Neatvideo has done it's job and before frameserving back into the MPEG2 encoder.

Now I'm wondering if the introduced videolag by Neatvideo is always the same or does it depend on the settings used for processing ?

Thanks,
Tobias (Germany)
NVTeam
Posts: 2745
Joined: Thu Sep 01, 2005 4:12 pm
Contact:

Post by NVTeam »

Yes, the lag depends on the filter settings. Specifically, it is equal to the Temporal filter radius.

Vlad
Fudoh
Posts: 7
Joined: Mon Jul 24, 2006 3:10 pm

Post by Fudoh »

Specifically, it is equal to the Temporal filter radius.
Thanks, good to know !

Tobias
crucci
Posts: 7
Joined: Fri May 05, 2006 2:37 am
Location: US

Frame serving from Virtual Dub using NV with temporal filter

Post by crucci »

Tobias,

I also am trying to frame serve from Virtual Dub and I am getting the audio lag error. How did you get around this? Did you disable the temporal filter, or did you somehow use Virtual Dub to resync the audio and video streams?

I am trying to frame serve into Ulead Video Studio 10. What program are you frame serving into?


thanks,
Chuck
Fudoh
Posts: 7
Joined: Mon Jul 24, 2006 3:10 pm

Post by Fudoh »

First I wanted to frameserve into CCE, but I skipped this since the Neat rendering time is so huge, that I want to check the result first before further working on the video, so I save to a HuffYUV File first now.

But I tried to correct the audio offset before and I worked fine. Just delay the audio by -1 frame (40ms in PAL) and it will be correct again.

You can also first render the audio and frameserve the video afterwards. Then in your video app (e.g. Ulead) simply load the audio file and do the delay there...

Tobias
Hawfinch
Posts: 3
Joined: Wed Aug 26, 2009 11:14 pm

Post by Hawfinch »

Just to be clear - does the temporal filtering always throw the audio out of sync in VirtualDub, or does this only happen when frameserving?

If I save as an AVI rather than frameserving, do I still need to resync the audio by adding a 40ms delay (if the Temporal filter radius = 1)?
NVTeam
Posts: 2745
Joined: Thu Sep 01, 2005 4:12 pm
Contact:

Post by NVTeam »

As far as I know, VirtualDub itself takes care about synchronizing the video and audio in situations when the video stream is delayed by a filter which introduces a lag of several frames. VirtualDub knows about the lag and corrects it. So, when saving to an AVI, I would expect no sync issues.
Hawfinch
Posts: 3
Joined: Wed Aug 26, 2009 11:14 pm

Post by Hawfinch »

Many thanks.

In fact, having just upgraded to VirtualDub 1.9.5 (the new stable release), I notice that I'm not getting the lag error message when frameserving to CCE. So perhaps this issue has been fixed now?

This is from the VDub 1.9.5 release notes:
Multi-frame fetching in video filters: The video filter system has been rewritten to support a request/process model. The result is that video filters can now explicitly request multiple source frames for each output frame they produce. This eliminates the need to do tricky internal buffering inside of a video filter to emulate such support and makes it much easier to implement video filters that require a frame window, such as field shifters and deinterlacers.
NVTeam
Posts: 2745
Joined: Thu Sep 01, 2005 4:12 pm
Contact:

Post by NVTeam »

It is good that the problem is fixed in VD but I think it is not the right quote. The fix should be described in some other words. The quote is about serving frames to filters but what we should look for is something about frame serving to CCE. Anyway, that problem seems to be fixed and that is good.
Post Reply